Scribd Logo
Загрузить

Добро пожаловать в Scribd!

  • T#

    Загружено:

    Om Prakash
    3 просмотров19 страниц
    Trading system

    Авторское право

    © © All Rights Reserved

    Доступные форматы

    PDF или читайте онлайн в Scribd

    Этот документ был вам полезен?

    Это неприемлемый материал?

    Пожаловаться на этот документ
    Trading system

    Авторское право:

    © All Rights Reserved
    Скачайте в формате PDF или читайте онлайн в Scribd
    Отметить как неприемлемый контент
    3 просмотров19 страниц

    T#

    Загружено:

    Om Prakash
    Trading system

    Авторское право:

    © All Rights Reserved
    Скачайте в формате PDF или читайте онлайн в Scribd
    Отметить как неприемлемый контент
    из 19
    Chapter 11 Trading System Many trading systems have been proposed by traders. However, theit usefulness is seldom analyzed. In this chapter, we will discuss how we can analyze a trading system mathematically, Some trading systems use multiple timeframes. We will discuss the advantages and disadvantages. One trading system, the triple screen trading system, will be taken as an ‘example 11.1 Multiple Timeframes Some traders recommend using several timeframes to analyze the market. A method of using triple screens has been suggested by Elder [1993, 2002]. In the first screen, a strategic decision to trade long or short is made using a trend-following indicator on a long-term chart. In the second sereen, a tactical decision about entries and exits is formed using oscillators on an intermediate-term chart, (Oscillators measure the speed or slope of the trend. Examples of oscillators are momentum, velocity, ete.) In the third screen, methods for placing buy and sell orders are implemented on a short-term chart, or on the same intermediate-term chart. The method can consist of, e.g., using a breakout or pullback to enter trades. An intermediate timeframe is any timeframe chosen by the trader. It can be a weekly chart, a daily chart, or a five-minute chart The long-term chart is decided by multiplying the unit time interval of the intermediate chart by a certain factor, e.g., five. If the intermediate chart is daily, then the long-term chart is weekly. If the intermediate chart is five minutes, then the long-term chart can be half-hourly. Similarly, the short-term chart is decided by dividing the unit time interval of the intermediate chart by the same factor. For a daily intermediate chart, the short-term chart can be hourly, For a five-minute 139 160 Mathematical Techniques in Financial Market Trading chart, the short-term chart is a one-minute chart. ‘The factor chosen is not that critical. It can be a factor of four instead of five. ‘Thus, if a trader's intermediate timeframe is a weekly chart, his long-term frame will be a monthly chart. What does it mean by going from an intermediate timeframe to a long-term timeframe? Let's take the factor of five. It simply means taking every fifth point of the data points in the intermediate time frame to form the data points in the long-term frame, Mathematically, it is called downsampling five [Mak 2003]. An original signal of frequency ‘2/10 radian will become a signal of m/2 radian in the long-term frame. A signal of 7/50 radian will become a signal of 7/10 radian. You can imagine the frequencies in the frequency response plot of the market price signal in the intermediate timeframe to stretch by a factor of five to form the frequencies in the frequency response plot in the long-term frame. Does a long-term frame help in analyzing the market? We will find out in the following section. 11.1.1 Long-Term Timeframe ILL.1.1 Advantages It is common for traders to apply a trending indicator, e.g, an exponential moving average (EMA) on the data, As said earlier, a trending indicator is a low pass filter. Compared to an exponential moving average of a certain length applied to the data in an intermediate time-chart, downsampling the data to a long-term time chart and applying the same exponential moving average would mean pushing more signals of high frequencies out of the original signal in the intermediate time frame. Signals of low frequencies will be more emphasized, This effectively means increasing the length of the exponential moving average in the intermediate time chart, thus narrowing the bandwidth of the low pass filter. However, there is an advantage of downsampling to a long-term chart as compared to increasing the length of an EMA in an intermediate chart. There is less phase lag. Mathematically, we would not be able to calculate the frequency response of an indicator on downsampled signal. However, we can Trading System 16 calculate the frequency response of a skipped convolution of an \dicator, which should provide some insight to an indicator on downsampled signal Fig 11.1(@) plots the amplitude response of an exponential moving average of length 130, and the amplitude response of a skipped 5 convolution of an exponential moving average of length 26, It can be seen that their amplitudes are approximately the same for @ < 0.5. That means, for @ < 0.5, the two EMA‘s have similar smoothing capability. It can actually be shown that, for @ close to 0, a skipped D convolution of fan exponential moving average of length M has similar smoothing capability to an exponential moving average of length approximately equal to DM, where D is the skip parameter. Fig 11.1(b) plots the phase response of an exponential moving average of length 130, and the phase response of a skipped 5 convolution of an exponential moving average of length 26. It can be seen that, except in the high frequency range, the skipped convolution has much less phase lag. This implies that the EMA of a downsampled signal has less phase lag than the phase of an EMA of similar smoothing capability in the original signal, ‘Thus, analyzing market data in a long-term timeframe does have fan advantage. The trending indicator would have less phase lag ‘compared to the same trending indicator with similar smoothing capacity in an intermediate timeframe. The long-term frame has also an obvious advantage that the trader has less data to handle and visualize. He can see the forest instead of the trees. However, it does have its disadvantages. 111.12 Disadvantages (2) While trending indicators, e.g., the EMA, provides less phase lag in downsampled signal than the original signal, oscillator indicators, e.g., the momentum indicator and velocity indicators provide more phase lag. We will illustrate # an example in the frequency domain. Fig 11.2(@) shows the amplitude of the frequency response of the cubic 162 Mathematical Techniques i Financial Market Trading spre meg aps Me (i: ged M= 26 0) 06 i os] rad Fesney ans Fig 11.1(a) The amplitudes of the Fourier Transform of an exponential moving average of length 130 (plotted as +) and a skipped 5 convolution of an exponential moving average of length 26 (plotted as x) are plotted versus the circular frequency © a Fig 1116) ‘Te phase ofthe Four Tranfom of an exponent moving eeagc OP tenn? 130 (lone a) ands cepped5 convolaton et 28 Sem at aking averse of lag 2 pot 3) ae ploted esi he Steularfequency © Trading System 168 ale ey deat (9), hed 0) cect Fegnney atin) * Fig 11.2(a) The amplitudes of the Fourier Transform of the cubie velocity indicator (plotted as +) and the skipped 5 convolution of the cubie velocity indicator (ploted asx) are plotted versus the circular frequency (a. ie eye (), sod a a Creda Fcurey ads) Fig 11.2(b) The phases of the Fourier Transform of the cubic velocity indicator (plotted as +) and the skipped 5 convolution of the cubic velocity indicator (plotted as x) are plotted versus the circular frequency «. 164 Mathematical Techniques in Financial Market Trading velocity indicator as well as the skipped 5 convolution of the cubic velocity indicator. Fig 11.2(b) shows the phase of the frequency response of the cubic velocity indicator as well as the skipped 5 convolution of the cubic velocity indicator. From Fig 11.2(b), it can be implied that except for frequency near r, cubic velocity on downsampled signal has a larger phase lag than the cubic velocity on the original signal We will also take a look at an example in the time domain. Fig 11,3(a) shows an original price signal having a frequency of n/16 radian (denoted by .) and the down 4 price signal which has a frequency of 7/4 (denoted by 0). ‘The 2-bar momentum indicator (which is the same as the 2-point moving difference) applied to the original signal has a phase lag of n/32 (= n/I6 x ¥4) radian from the ideal phase lead of n/2 [Mak 2003] ‘The output is shown in Fig 11.3(b). The unit impulse response of the ‘momentum indicator is defined here as (1,1). As it takes two points to calculate the momentum, the very first point (which is arbitrarily set to zer0) in Fig 11.3(b) should be ignored. Fig 11.3(c) shows the down 4 ‘momentum signal. It can be seen that the phase difference between the

    Вам также может понравиться